Detection on T-SROIE dense text for robustness (test)
99.9AccuracyGemini-2.5-Pro
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Gemini-2.5-Pro2026.04 | 99.9 | 99.9 | |
| qwen3-vl-8B2026.04 | 99.9 | 99.9 | |
| DocShield2026.04 | 98.8 | 99.4 | |
| InternVL-3.5-8B2026.04 | 86.4 | 92.7 | |
| gpt-4o-o3-04162026.04 | 84.5 | 91.6 | |
| FakeShield2026.04 | 14.7 | 25.7 | |
| DeepSeekVL-7B2026.04 | 11.3 | 20.3 | |
| Qwen2.5-VL-7B2026.04 | 6.1 | 11.5 |